A new robust model-free feature screening method for ultra-high dimensional right censored data

Yi Liu and Xiaolin Chen

Communications in Statistics - Theory and Methods, 2022, vol. 51, issue 6, 1857-1875

Abstract: This paper is concerned with the robust feature screening method for ultra-high dimensional right censored data. A new robust and model-free feature screening approach is built upon a screening index constructed from a fresh measure of dependence between the survival time and a single covariate. One attractive property of this newly introduced index is that it equals zero if the survival time and covariate are independent. In addition, it is invariant to monotonic transformation of the response, and robust against heavy-tailed distributions and outliers for the response. These appealing properties render the derived feature screening approach to be a competitive one among feature screening methods for ultra-high dimensional right censored data. We establish the sure screening property for the suggested feature screening procedure under some regularity conditions, and evaluate its performance through numerical studies. Numerical comparisons with several main competitors show the advantages of the newly proposed means over its main competitors. The well-known diffuse large-B-cell lymphoma (DLBCL) data is utilized to illustrate our methodology.

Date: 2022
References: Add references at CitEc
Citations:

Downloads: (external link)
http://hdl.handle.net/10.1080/03610926.2020.1769672 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:lstaxx:v:51:y:2022:i:6:p:1857-1875

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/lsta20

DOI: 10.1080/03610926.2020.1769672

Access Statistics for this article

Communications in Statistics - Theory and Methods is currently edited by Debbie Iscoe

More articles in Communications in Statistics - Theory and Methods from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().